The article still applies to the things which have replaced Hadoop. The operational, development, and computational overhead of distributing/re-aggregating work remains huge.